DOS ain't dead

Hi,

did anyone succeed to run the LFN-aware version of rxdos (source files dated 08.08.99)? I'm unable to run this version, neither on a real machine nor in Qemu. On the real hardware - with a FAT32 drive - it just displays multiple times "booting rxdos" and freezes. In Qemu with just a FD device and no HDs, I get a A:\> prompt, but all what works is VER and CD, no possiblility to start an external program or running DIR.

EDIT by rr: moved to correct category, changed subject from "rxdos"

---
MS-DOS forever!

I cannot test RxDOS now, but I tried it ca 8 años ago and I had the same problems on real hardware.

> did anyone succeed to run the LFN-aware version of rxdos (source files
> dated 08.08.99)? I'm unable to run this version, neither on a real machine
> nor in Qemu. On the real hardware - with a FAT32 drive - it just displays
> multiple times "booting rxdos" and freezes. In Qemu with just a FD device
> and no HDs, I get a A:\> prompt, but all what works is VER and CD, no
> possiblility to start an external program or running DIR.

Works fine on Virtual PC 2004, although RXD_BOOT.COM refuses to work on Windows (2000). I had to boot MS-DOS in Virtual PC and run MAKEBOOT.BAT from there.

---
Forum admin

> Works fine on Virtual PC 2004, although RXD_BOOT.COM refuses to work on
> Windows (2000). I had to boot MS-DOS in Virtual PC and run MAKEBOOT.BAT
> from there.

Thanks, it "works" with VPC 2007 for me as well. With "works", I mean the DIR command displays something and the prompt returns, which is ways more than what I got with Qemu & Bochs.

---
MS-DOS forever!

> did anyone succeed to run the LFN-aware version of rxdos (source files
> dated 08.08.99)?

Yes it works fine for me. (Although RxDOSCMD has some command line editing bugs and not all partitions are logged in.) The 1999-08-08 source files are version 7.1.5 from ibiblio.org/FreeDOS mirrors. (Though the binaries (rxdos-exes.zip) are dated 1999-08-24.)

> I'm unable to run this version, neither on a real machine
> nor in Qemu. On the real hardware - with a FAT32 drive - it just displays
> multiple times "booting rxdos" and freezes.

Do you have "non-DOS" (either logical or primary) partitions on this machine? ("non-DOS" is anything except FAT, FAT32, Extended. Hidden partitions are also "non-DOS".) Although the problem doesn't occur on my DOS machine info.txt on ibiblio.org states they're problems booting RxDOS if it finds "non-DOS" partitions.

Anyway, the FAT32 support of this RxDOS version seems rather good. It didn't corrupted my 7 GiB FAT32 drive when I tested it. (The info.txt reports a bug on >2GiB drives but I didn't see it yet.)

> In Qemu with just a FD device
> and no HDs, I get a A:\> prompt, but all what works is VER and CD, no
> possiblility to start an external program or running DIR.

Sounds like running RxDOSCMD without LFN services... Something blocking file services for RxDOSCMD. (It doesn't have any help or error messages yet, so DIR just displays nothing if it got an error.)

> it "works" with VPC 2007 for me as well. With "works", I mean the
> DIR command displays something and the prompt returns, which is ways more
> than what I got with Qemu & Bochs.

Actually RxDOS isn't really usable yet

---
l

> Do you have "non-DOS" (either logical or primary) partitions on this
> machine? ("non-DOS" is anything except FAT, FAT32, Extended. Hidden
> partitions are also "non-DOS".) Although the problem doesn't occur on my
> DOS machine info.txt on ibiblio.org states they're problems booting RxDOS
> if it finds "non-DOS" partitions.

It runs in VPC, that's sufficient for now.

> > In Qemu with just a FD device
> > and no HDs, I get a A:\> prompt, but all what works is VER and CD, no
> > possiblility to start an external program or running DIR.
>
> Sounds like running RxDOSCMD without LFN services... Something blocking
> file services for RxDOSCMD. (It doesn't have any help or error messages
> yet, so DIR just displays nothing if it got an error.)

No, LFN services are active. In VPC, if I run "mkdir" (without param), I get 2 FAT entries, first an LFN one, then a "short" one (both contain garbage, though). In Qemu, "dir" doesn't just display nothing, it "hangs".

> Actually RxDOS isn't really usable yet

After the first rather bad impression it's beginning to make fun. I can load JemmEx from the commandline ("jemmex noinvlpg load"), hdpmi32 works and - after a small bugfix concerning int 21h, ah=0Ah - FD debug.com/debugx.com work as well.

---
MS-DOS forever!

> No, LFN services are active. In VPC, if I run "mkdir" (without param), I
> get 2 FAT entries, first an LFN one, then a "short" one (both contain
> garbage, though). In Qemu, "dir" doesn't just display nothing, it
> "hangs".

I meant the error looked like one was trying to run RxDOSCMD on non-LFN systems. Because RxDOSCMD relies on the LFN services (by now) it won't work on these systems, resulting in i.e. DIR and all external programs not working.

> After the first rather bad impression it's beginning to make fun. I can
> load JemmEx from the commandline ("jemmex noinvlpg load"), hdpmi32 works
> and - after a small bugfix concerning int 21h, ah=0Ah - FD
> debug.com/debugx.com work as well.

I won't call it an usable MS-DOS alternative because of this, though.

---
l

> I won't call it an usable MS-DOS alternative because of this, though.

Well, there seems to still be confusion about which version is the latest or not. (BTW, I don't think Mike P. ever did update Sourceforge. At least, it doesn't look like it.) Is this newest?? (I'm pretty sure this is the A86 port.)

(RXDOS_BINARIES_7.1.ZIP):

RXDOS.SYS      25,734        (41,390)          Jan,25,2000  01:38:48p
RXDOSCMD.COM   10,289        (32,774)          Jan,25,2000  12:23:46p
RXDOSBIO.SYS   4,117         (6,723)           Jan,10,2000  06:38:34p
RXD_BOOT.COM   1,114         (4,684)           Jan,05,2000  08:47:42p
SYS.BAT        52            (79)              Jan,26,2000  10:10:58a

> Well, there seems to still be confusion about which version is the latest
> or not. (BTW, I don't think Mike P. ever did update Sourceforge. At least,
> it doesn't look like it.) Is this newest?? (I'm pretty sure this is the A86
> port.)
>
> (RXDOS_BINARIES_7.1.ZIP):
>
> RXDOS.SYS      25,734        (41,390)          Jan,25,2000  01:38:48p
> RXDOSCMD.COM   10,289        (32,774)          Jan,25,2000  12:23:46p
> RXDOSBIO.SYS   4,117         (6,723)           Jan,10,2000  06:38:34p
> RXD_BOOT.COM   1,114         (4,684)           Jan,05,2000  08:47:42p
> SYS.BAT        52            (79)              Jan,26,2000
> 10:10:58a

posting binaries with date won't help much, just sources are interesting.

There's also no guarantee that the "newest" things are the "best" ones. If someone looses interest in his project, he won't care anymore about bugs, docs and code quality.

---
MS-DOS forever!

> Well, there seems to still be confusion about which version is the latest
> or not. (BTW, I don't think Mike P. ever did update Sourceforge. At least,
> it doesn't look like it.)

Yes. It seems on rxdos.SF.net are some RxDOS 6.22? source files and RxDOS 7.00? binaries or similar.

> Is this newest?? (I'm pretty sure this is the A86 port.)

The A86 port has the version number a7.1.5. I think it's almost equal to the MASM version 7.1.5, which is the one on the FreeDOS mirrors. RXD_BOOT is better in the A86 port - no unused padding, .COM instead of .EXE file. RxDOSCMD is also a .COM file but I don't know if it changed otherwise. I think the A86 port doesn't have debug information inside the shipped binaries. (The binaries on the FreeDOS mirrors have CodeView information appended.)

---
l

> The A86 port has the version number a7.1.5. I think it's almost equal to
> the MASM version 7.1.5, which is the one on the FreeDOS mirrors. RXD_BOOT
> is better in the A86 port - no unused padding, .COM instead of .EXE file.
> RxDOSCMD is also a .COM file but I don't know if it changed otherwise. I
> think the A86 port doesn't have debug information inside the shipped
> binaries. (The binaries on the FreeDOS mirrors have CodeView information
> appended.)

I never noticed. But it's true, A86 doesn't add debug symbols except in separate .SYM files (for D86).

> After the first rather bad impression it's beginning to make fun.

So, will we be seeing a JWasm compilation?

> > After the first rather bad impression it's beginning to make fun.
>
> So, will we be seeing a JWasm compilation?

That's an old hat already : rxdos + jwasm

However, I'm still not sure if RxDos deserves resurrection. The feature which makes it somewhat interesting is the build-in support for LFN, but exactly this feature is broken - or, more likely, did never work.

---
MS-DOS forever!

> However, I'm still not sure if RxDos deserves resurrection. The feature
> which makes it somewhat interesting is the build-in support for LFN, but
> exactly this feature is broken - or, more likely, did never work.

Yes. If you want DOS with build-in LFN take Datalight ROM DOS

---
DOS-u-akbar!

> > However, I'm still not sure if RxDos deserves resurrection. The feature
> > which makes it somewhat interesting is the build-in support for LFN,
> but
> > exactly this feature is broken - or, more likely, did never work.
>
> Yes. If you want DOS with build-in LFN take Datalight ROM DOS

Actually I can't find the (free or not) source of ROM-DOS so I'll still try to get RxDOS's LFN functions working.

Anyway the LFN functions of RxDOS worked very well for me sometimes - but sometimes it also just appends garbage to the directory as Japheth reported. I hope they're just some bugs preventing it from working properly.

---
l

> Actually I can't find the (free or not) source of ROM-DOS so I'll still
> try to get RxDOS's LFN functions working.

ROM-DOS source code is not free. So don't waste your time searching.

---
Forum admin